Ask how much creatine people use and at least three different answers come back, all of them sourced, all of them real. They are not versions of one number that has been garbled. They come from three different kinds of document — a biopsy experiment, a professional recommendation, and the arithmetic mean of what researchers actually handed out. This page separates them. It is research journalism, reports amounts with their sources and dates, and contains no usage guidance, per our editorial standards.
Number one: 20 g for six days, then 2 g a day — the measurement
This is the oldest and the most precisely established, and it comes from a single paper. In 1996, Hultman and colleagues studied 31 men ingesting creatine in different quantities over different periods, with muscle biopsies and urinary creatinine measured throughout.
What the paper reports:
- Muscle total creatine rose about 20% after 6 days at 20 g/day.
- That elevated concentration was maintained at 2 g/day for a further 30 days.
- Without the 2 g/day, total creatine declined gradually, and 30 days after stopping it was no different from the pre-supplementation value.
- A rate of 3 g/day produced a similar 20% rise over 28 days.
(Hultman et al., J Appl Physiol 1996;81(1):232–7, PMID 8828669; doi:10.1152/jappl.1996.81.1.232)
The paper's own conclusion contains the sentence most loading advice leaves out: 3 g a day "is in the long term likely to be as effective at raising tissue levels as this higher dose". The fast route and the slow route arrive at the same place. Our page on what the loading trials measured follows that finding through both founding studies, including the fact that neither of them measured strength or muscle size at all.
Number two: 3 to 5 g a day — the recommendation
This is the figure in general circulation, and it is a professional recommendation rather than a direct measurement. The 2017 International Society of Sports Nutrition position stand sets out the underlying accounting: about 1 to 2% of intramuscular creatine degrades to creatinine daily, so the body replaces roughly 1 to 3 g a day to hold normal unsupplemented stores, with about half of that coming from food. On top of that baseline the same document records 5 to 10 g/day for larger athletes in intense training and 10 to 30 g/day for clinical populations with synthesis deficiencies or therapeutic aims (Kreider et al., J Int Soc Sports Nutr 2017, PMID 28615996).
So the familiar small number is a maintenance figure with margin added for body size, and the same document carries figures five to ten times larger for other populations. The reason the figures scale is on what creatine is: the store has a fixed ceiling per kilogram of muscle, so a larger muscle mass is a larger tank.
Number three: 12.5 g a day — what the trials actually administered
This is the number nobody quotes, and it is the one that most changes how trial results should be read.
The 2025 safety analysis assembled 685 human clinical trials of creatine supplementation. Reporting what those trials gave people, it records that 95% used creatine monohydrate at an average of 0.166 g/kg/day — about 12.5 g/day — for a mean of 64.7 days, in studies running up to 14 years (Kreider et al., J Int Soc Sports Nutr 2025, PMID 40198156; doi:10.1080/15502783.2025.2488937).
| Figure | What kind of number it is | Source |
|---|---|---|
| 20 g/day × 6 d, then 2 g/day | Measured biopsy protocol | Hultman 1996 |
| 1–3 g/day baseline; 3–5 g commonly quoted; 5–10 g larger athletes; 10–30 g clinical | Professional recommendation | ISSN position stand 2017 |
| 0.166 g/kg/day ≈ 12.5 g/day for 64.7 days | Arithmetic mean of what 685 trials administered | Kreider 2025 |
The consequence is worth stating carefully. When a study is cited as evidence that creatine does something, the amount behind that evidence averaged around 12.5 g a day — two to four times the figure most often printed as a daily amount. That does not mean the smaller figure fails to maintain a full store; the 1996 data say it does. It means the outcome literature and the maintenance recommendation are not measuring the same exposure, and quoting a trial result alongside a maintenance figure silently swaps one for the other.
What the surplus does
Large amounts do not accumulate indefinitely, and the 1992 study measured where they go. In three subjects receiving six 5 g doses a day, renal excretion accounted for 40%, 61% and 68% of the administered creatine over the first three days, while muscle uptake was concentrated in the first two days and accounted for 32% of the dose in those subjects (Harris et al., Clin Sci 1992, PMID 1327657).
That is the ceiling working. Muscle total creatine averages about 120 mmol per kilogram of dry muscle mass against an upper limit of about 160, so the space available is finite and the rest is filtered out. It is also the mechanism behind a blood-test artefact that causes real confusion, which we separate on what creatine looks like in bloodwork and on creatine and the kidneys.
Two places where the amount demonstrably changed the result
Amount is not a cosmetic detail in this literature — there are documented cases where it decided the finding.
Cognition. The EFSA nutrition panel, refusing a cognitive-function health claim in November 2024, made the amount central to its reasoning: the acute working-memory effect appeared at 20 g/day for 5 to 7 days and was not observed at 2.2 to 14 g/day, nor with continuous consumption of 5 g/day for six weeks after a five-day loading phase (EFSA Journal 2024;22(11):e9100). An effect that appears only at one exposure and vanishes on either side of it is the kind of inconsistency regulators weigh heavily. The full grading is on creatine benefits by endpoint.
Who responds. A 2017 spectroscopy study administered a standardised 0.3 g/kg/day for 7 days to children, omnivorous adults, vegetarians and elderly participants. The rise in muscle phosphocreatine was statistically significant in vegetarians, children and the elderly, and did not reach significance in omnivorous adults (p = 0.3348) (Solis et al., J Appl Physiol 2017, PMID 28572496). The authors' conclusion is that a single universal protocol "may lead to heterogeneous muscle responses in different populations".
A note on one source this page does not cite
The NIH Office of Dietary Supplements fact sheet is the document AI assistants most often quote for creatine amounts. It could not be retrieved by any automated request on 2026-09-16 — every attempt returned an HTTP 403 interstitial — so it is not quoted here, and no figure on this page is attributed to it. The primary studies above are the sources that fact sheet itself summarises, and they were opened directly.
The honest summary
- 20 g/day for six days then 2 g/day is a measured biopsy protocol from 1996, and the same paper found 3 g/day reaches the same concentration in 28 days.
- 1 to 3 g/day is the replacement requirement; 3 to 5 g is the common conversion of it; 5 to 10 g and 10 to 30 g appear in the same professional document for larger athletes and clinical populations.
- 12.5 g/day is the mean of what 685 trials actually administered, which is the exposure behind most published creatine results.
- The store ceiling is about 160 mmol/kg dry muscle; beyond it, the surplus is excreted, and 40 to 68% of a loading dose left in urine within three days in the founding study.
This page reports amounts published in trials, position statements and regulatory opinions, each with its source and date. It is not advice, it recommends nothing, and decisions about supplementation belong with a clinician.
Sources
- Hultman E, Söderlund K, Timmons JA, Cederblad G, Greenhaff PL. Muscle creatine loading in men. J Appl Physiol 1996;81(1):232–7 — PMID 8828669 · doi:10.1152/jappl.1996.81.1.232
- Harris RC, Söderlund K, Hultman E. Elevation of creatine in resting and exercised muscle of normal subjects by creatine supplementation. Clin Sci 1992;83(3):367–74 — PMID 1327657 · doi:10.1042/cs0830367
- Kreider RB, Gonzalez DE, Hines K, Gil A, Bonilla DA. Safety of creatine supplementation: analysis of the prevalence of reported side effects in clinical trials and adverse event reports. J Int Soc Sports Nutr 2025;22(sup1):2488937 — PMID 40198156 · doi:10.1080/15502783.2025.2488937
- Kreider RB, Kalman DS, Antonio J, et al. ISSN position stand: safety and efficacy of creatine supplementation in exercise, sport, and medicine. J Int Soc Sports Nutr 2017;14:18 — PMID 28615996
- EFSA NDA Panel. Creatine and improvement in cognitive function. EFSA Journal 2024;22(11):e9100 — PMID 39564533 · doi:10.2903/j.efsa.2024.9100
- Solis MY, Artioli GG, Otaduy MCG, et al. Effect of age, diet, and tissue type on PCr response to creatine supplementation. J Appl Physiol 2017;123(2):407–14 — PMID 28572496 · doi:10.1152/japplphysiol.00248.2017
All sources retrieved and opened 2026-09-16. Digital object identifiers verified by content negotiation on the same date.
